Releases a public trace of coding-agent LLM sessions and characterizes workload features including long loops, context lengths, tool diversity, and cache hit rates for serving optimization.
Speed at the cost of quality: How cursor ai increases short-term veloc- ity and long-term complexity in open-source projects, 2026
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
TraceLab: Characterizing Coding Agent Workloads for LLM Serving
Releases a public trace of coding-agent LLM sessions and characterizes workload features including long loops, context lengths, tool diversity, and cache hit rates for serving optimization.